Re: [PATCH 1/3] btrfs: add readmirror type framework

[Anand Jain <anand.jain@oracle.com>]

On 12/20/19 10:44 PM, Josef Bacik wrote:
> On 12/20/19 12:06 AM, Anand Jain wrote:
>> As of now we use %pid method to read stripped mirrored data. So
>> application's process id determines the stripe id to be read. This type
>> of routing typically helps in a system with many small independent
>> applications tying to read random data. On the other hand the %pid
>> based read IO distribution policy is inefficient if there is a single
>> application trying to read large data and the overall disk bandwidth
>> remains under utilized.
>>
>> So this patch introduces a framework where we could add more readmirror
>> policies, such as routing the IO based on device's waitqueue or manual
>> when we have a read-preferred device or a policy based on the target
>> storage caching.
>>
>> Signed-off-by: Anand Jain <anand.jain@oracle.com>
>> ---
>>   fs/btrfs/volumes.c | 16 +++++++++++++++-
>>   fs/btrfs/volumes.h |  8 ++++++++
>>   2 files changed, 23 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
>> index c95e47aa84f8..0c6caae29248 100644
>> --- a/fs/btrfs/volumes.c
>> +++ b/fs/btrfs/volumes.c
>> @@ -1162,6 +1162,8 @@ static int open_fs_devices(struct 
>> btrfs_fs_devices *fs_devices,
>>       fs_devices->opened = 1;
>>       fs_devices->latest_bdev = latest_dev->bdev;
>>       fs_devices->total_rw_bytes = 0;
>> +    /* Set the default readmirror policy */
>> +    atomic_set(&fs_devices->readmirror, BTRFS_READMIRROR_DEFAULT);
> There's no reason for this to be atomic, it's just a behavior change, if 
> you really want to be super safe use READ_ONCE/WRITE_ONCE and have 
> readmirror be your enum.  Thanks,

  Agreed fs_devices::readmirror doesn't have to be atmoic_t. Fixed this
  to declare it as u8 in v2.

Thanks, Anand


> Josef